Okta AI Agent

The Okta AI Agent is an innovative tool designed specifically for Okta administrators, allowing them to query tenant data using natural language. This AI-powered assistant is ideal for IT administrators, Identity and Access Management (IAM) professionals, GRC (Governance, Risk, and Compliance) teams, and auditors. It leverages advanced enterprise AI models to simplify interactions with Okta data, transforming complex queries into straightforward language.

Key Features

• Natural Language Queries: Interact with your Okta data in simple English, making it accessible for users without extensive technical knowledge.
• Multiple AI Providers: Benefit from various AI services, including:
  • Google Vertex AI (Gemini 1.5 Pro)
  • OpenAI (GPT-4)
  • Azure OpenAI (GPT-4)
  • Ollama, offering local, self-hosted AI capabilities
• Easy Okta Sync: Effortlessly synchronize Okta data to a local SQLite database for offline access and analysis.

Vision for the Future

The long-term goal of the Okta AI Agent is to evolve it into a fully autonomous agent capable of executing nearly all Okta administrative functions without compromising on security or compliance.

Security and Privacy

The architecture ensures that security is paramount, with all Okta data being stored locally. Users maintain control over their Okta API tokens, and sensitive organizational data remains protected, as nothing is sent to language models apart from user prompts and system messages.

Data Model Overview

The agent supports querying various entities, including users, groups, and applications, with relevant fields and relationships clearly outlined:

• Users: Attributes such as email, login, and department are captured.
• Applications: Contains details like name, status, and sign_on_mode.
• Groups: Includes descriptive fields for better management.
• Policies: Captures various policy-related metrics.

Getting Started

To utilize the Okta AI Agent, users are guided through a straightforward setup process involving cloning the repository, configuring the environment, and running data synchronization scripts. Sample queries help users quickly adapt to the system, enabling interactions such as:

python Fetch_Okta_Data.py
python Okta_AI_Agent.py

Community Engagement

The project encourages contributions and improvements from the community. Users can share queries, explore community submissions, and participate in ongoing discussions about enhancements and features.

By adopting the Okta AI Agent, organizations can streamline their Okta management processes, ensuring efficiency, security, and ease of use.